From a YouTube Video to an Anki Deck in One Sitting

Every YouTube video can carry one or more caption tracks. Creators can upload subtitle files or type captions in YouTube Studio; YouTube also generates automatic captions using speech recognition for many languages. Both are stored as timed text: each cue has a start time, an end time, and the text spoken. Nativra reads these tracks and uses the timings to highlight the current line as the video plays and to stamp each saved word with the moment it was said.

The distinction between creator-uploaded and auto-generated captions matters a lot for language study. Auto-generated captions are good for clear, single-speaker speech in major languages and poor for overlapping dialogue, strong accents, music, and smaller languages. They also lack punctuation and sentence boundaries, which makes the sentence field of a card less useful. Prefer videos with uploaded captions when you can.

The best videos for card mining share a few properties: a single clear speaker or well-separated dialogue, a topic you already know something about (so context helps comprehension), a length of five to fifteen minutes (so a session produces a manageable deck), and vocabulary density slightly above your level. Nativra's CEFR grading shows the level distribution of a video's captions before you commit to it.

Genres that work well: interviews, explainers, cooking and DIY channels, news summaries, and vlogs. Genres that are harder: comedy (wordplay, speed), music, and multi-person panel shows. Start easy; the point is volume of comprehensible input, not proving you can survive a hard video.

• Prefer creator-uploaded captions over auto-generated.
• Single speaker, clear audio, 5–15 minutes.
• Topic you have background knowledge in.
• CEFR distribution mostly at or one level above yours.

Nativra exports a tab-separated file with one line per saved word: word, lemma, sentence, meaning, CEFR level, video title, and timestamp. In Anki, File → Import, choose the file, confirm the tab delimiter, choose a deck and note type, and map the columns. Tag with the video title so you can suspend or study a video's cards as a group. If you re-export the same video after saving more words, Anki updates existing notes by first field rather than duplicating, provided the word column stays first.

You can also skip Anki entirely and review inside Nativra, which schedules cards with the SM-2 algorithm. Export when you want a single place for all your decks or want Anki's mobile apps offline.

Nativra fetches captions; it does not download or store the video, and it cannot produce captions for a video that has none. Some videos restrict caption access, and live streams do not have caption tracks until they are processed. For those cases, an SRT or VTT file from another source can be uploaded instead. Nativra does not extract audio clips for cards; the timestamp is the link back to the sound.